Classifying Players For Unique Game Experiences

togelius writes "Whenever you play a game of Tomb Raider: Underworld, heaps of data about your playing style is collected at Eidos' servers. Researchers at the Center for Computer Games Research have now mined this data to identify the different types of player behavior (PDF). Using self-organizing neural networks, they classified players as either Veterans, Solvers, Pacifists or Runners. It turns out people play the game for very different reasons and focus on different parts of the game, but almost everyone falls into one of these categories. These neural networks can now quickly determine which of these groups you belong to based on just seeing you play. In the near future, such networks will be used to adapt games like Tomb Raider while they are played (e.g. by removing or adding puzzles and enemies), so you get the game you want."

Comments

Foruc on different parts of game

[sopssa]

It turns out people play the game for very different reasons and focus on different parts of the game, but almost everyone falls into one of these categories.

Yep, I've noticed this too. I dont get why, but some people tend to stare the ass more, while personally I like to enjoy the boobs.

Did this research notice if there were any deaths caused by getting discracted when you jumped and the camera got into such position that you tried to get a nippleslip or see the panties?

Re:Foruc on different parts of game

[ShieldW0lf]

Just what we need... surround ourselves with ourselves. That will challenge us and cause us to grow into intelligent, tolerant and well rounded individuals.

Re:Foruc on different parts of game

[Andr T.]

Did this research notice if there were any deaths caused by getting discracted when you jumped and the camera got into such position that you tried to get a nippleslip or see the panties?

Just what we need... surround ourselves with ourselves. That will challenge us and cause us to grow into intelligent, tolerant and well rounded individuals.

/. just keep getting weirder and weirder.

Re:Foruc on different parts of game

[Minwee]

Just what we need... surround ourselves with ourselves. That will challenge us and cause us to grow into intelligent, tolerant and well rounded individuals.

I don't like hearing this kind of talk, so I'm going to mod it down.

No thank you

[RemoWilliams84]

After seeing how Tivo and Netflix recommendations go sometimes, I'm not sure I want a game changing itself because it thinks I know what I want. Not to knock Tivo or Netflix, they are accurate alot, but sometimes they are way off base.

Besides, if it knew what I really wanted, everything would just end up having tits.

Re:No thank you

[Anonymous Coward]

Besides, if it knew what I really wanted, everything would just end up having tits.

So you've never played Tomb Raider, then, have you?

Thanks for the heads up

[NervousNerd]

Whenever you play a game of Tomb Raider: Underworld, heaps of data about your playing style is collected at Eidos' servers.

Thanks for the heads up, so I won't buy it. I personally don't like having everything I do monitored in some way on some server with a shady privacy policy.

Re:Thanks for the heads up

[J_DarkElf]

Then don't buy the Xbox version. If you RTFA, it mentions that the data collection was done through Xbox Live.

Of course with its achievements etc. Xbox Live is always tracking everyone in the first place, Eidos' data collection is a logical next step. If you're paranoid, avoid Xbox Live, PSN, and any similar system (including Steam on PC unless firewalled).

Or of course just pull the network plug of the PC or console...

Re:Thanks for the heads up

[mcvos]

Totally. This is yet another attempt by the government to monitor its drones and keep them in line, another little teeter down the slippery slope to an Orwellian future.

But by analyzing how you react to this, they'll be able to offer you the customised Orwellian future that you really want.

Re:Thanks for the heads up

[Applekid]

Unless you're planning to write your social security number with bullet holes in the wall, I think you might be overreacting.

Using my play data to serve ads? No, thanks, I'll pass. Using my play data to realize I hate having to kill things in Tomb Raider? Sounds like a win to me.

Re:Thanks for the heads up

[SL Baur]

I think they are, though I'm sure that they're a little overwhelmed by the amount of data involved.

In the most recent patch 3.2 they removed "twinks" from regular battlegrounds and added XP. The vast majority of us cheered.

Re:Thanks for the heads up

[dkleinsc]

I think you guys have the wrong dystopia here: This isn't an Orwellian future that this sort of thing leads to, it's more of a Brave New World with perfectly customized soma for you.

Does it take nudity into account?

[broknstrngz]

How about the naked Lara Croft modders? Which slot do they fall into?

Re:Does it take nudity into account?

[Anonymous Coward]

Solvers: Solved her tight clothing getting in the way.

Re:Does it take nudity into account?

[Andr T.]

How about the naked Lara Croft modders? Which slot do they fall into?

Tricky question, don't you think? /. is a family-friendly website and nobody should answer that question.

(Insert 'you must be new here' joke now)

Bartle did this work already

[HBI]

...15 years ago. They change the names and claim it as unique research?

Re:Bartle did this work already

[am 2k]

...15 years ago. They change the names and claim it as unique research?

No. Bartle's taxonomy is only really relevant for MMORPGs and MUDs. This one is mostly for first person shooters and similar games.

Re:Bartle did this work already

[am 2k]

Here. But you could have found that yourself on Wikipedia...

Missing player type - metagamers

[petterb]

They seem to have forgotten about the metagamers :)

Almost everyone?

[noname444]

Using self-organizing neural networks, they classified players as either Veterans, Solvers, Pacifists or Runners ... but almost everyone falls into one of these categories

I didn't RTFA but wouldn't everyone fall into one of the categories? I mean, it sounds like the system does just that: puts the player in one of the categories.

Re:Almost everyone?

[jtogel]

The categories did not exist prior to the data; they were found by unsupervised learning algorithms in the data.

Re:Almost everyone?

[Sockatume]

The system discovers the categories. The analysis finds groupings of players who behave in similar ways through the game, and the researchers named those after-the-fact. There's no a priori reason why the players should group at all, though - the study could've equally found that only a small percentage of players clustered and the majority were radically different from each other.

Re:Almost everyone?

[zacronos]

Maybe it's someone who just like to run straight into a crowd of enemies and immediately die repeatedly for hours on end.

Runner

Or someone who logs in and just sits there not moving for hours. Or any number of other things that probably wouldn't even be classified as "play"

Pacifist or Solver

See, that's you doing it backwards -- once the categories have been defined by the neural networks (and labeled after the fact by humans), you are now trying to take any given data point and fit it into one of the categories. That's not how it works. Imagine looking at a 2D image containing many dots; if you were asked to draw perimeters around any significant clusters, you could probably do so without difficulty -- but depending on the 2D image you are given, it is entirely possible (even probable) that not every dot is going to be part of a cluster. If you wanted to include every dot, you could instead subdivide the image into regions, but that is a different task.

To speak to the specific examples, if Runner has been defined as something akin to "someone who likes to achieve the game's goals in as short a period as possible, skipping past subgoals and rewards if they present significant risk or slowdown", then in that case "someone who just like to run straight into a crowd of enemies and immediately die repeatedly for hours on end" could not be defined as a Runner. Similarly, if a Pacifist is defined as "someone who attempts to accomplish the game's goals using the minimum amount of violence possible", and Solver is defined as "someone who attempts to accomplish every task presented in the most efficient way possible, even when such tasks are not necessary to progress in the game", then "someone who logs in and just sits there not moving for hours" could not be defined as a Solver or a Pacifist. If you want to assign one of the labels to any given player description, sure you can do that, but it doesn't mean that player is actually part of the data cluster which defines the category.

Re:Almost everyone?

[gnick]

But, once you've run your data through and decided that 4 categories are sufficient, most designers (including myself) will restrict the NN to those categories. And somebody with really weird behavior will get lumped in and will slightly skew the existing category. The guy who runs into a crowd and dies over and over again may be described as a Runner, but he'll be an outlier in the runner class and his behavior will tweak the definition of a Runner.

Your options are to ignore outliers like him to avoid polluting your class, add a new class for people with that kind of behavior if there are enough of them to justify it, or (most likely) just accept that outliers skew tight groups and lump him in as a Runner - If the group is tight enough and he's rare enough, it won't matter.

Ideally, however, your architecture will be flexible enough that you can weigh how good a fit each player is to each group and adjust accordingly. I.e. adjust every obstacle according to a best-fit weighting rather than just delivering 4 different options on each level. Not having played the game or reading TFA, I can't speculate on that front.

So no variety?

[bcmm]

Many players enjoy some variety within a game. I've played all the Hitman games with the aim of completing the missions "cleanly", so I enjoyed the ones which force you to play the last mission as more of a shooter game (they did this in the 1st, 2nd and 4th games, while the third had a finale which offed the chance to play stealthily, but was still designed to produce a massive firefight if not played stealthily).

I would be somewhat annoyed if Eidos based the style of the final level of the next Hitman game on stats from the rest of the game, which seems to be a real possibility since Hitman is a game which offers plenty of chances to choose between stealth and action gameplay.

Re:So no variety?

[bigngamer92]

It would be interesting to see this in a strategy game like Civ 4. If you spend all your time in economics then the game will ease off the aggresive AI.

Of course just making the AI better would help a lot.

So the game is spyware?

[WCMI92]

I don't like the idea of BUYING something and then having my use of it monitored. That's no different than spyware.

Re:So the game is spyware?

[Rogerborg]

Sweet Zombie Jesus, the tin foil hat brigade are out in force today. The game is already awarding you Achievements as you play. You don't like being "spied" on to earn Achievements? Then why are you playing on XBox Live?

Oh, you didn't realize that this only applies to the XBox Live version? You didn't even read the article, you say? I've just earned the "Shocked and Stunned" Achievement.

Re:So the game is spyware?

[AP31R0N]

Ahhh! They're after you!

Spyware watches you to target advertising at you, and to help companies figure out how to optimize their costs and profits. This game is watching you so they can make games more appealing to more players. A game that designed to appeal to one play style will likely annoy the other types. Your Solver will complain about the lack of puzzles or over dependence on violence. If you can make a game cater to multiple styles, more people will speak well of it and more people will want the sequel. Yeah, you're helping them gain some competitive edge over another company, but you're also likely to get better games. Or games with broader appeal, without losing the niche players.

If this game can sense that i don't dig puzzles, it could send me more bad guys to kill. It could create opportunities for less violent play by letting me sneak around or by negotiating with NPCs. If it can do this w/o me thinking about it, so much the better. i don't want to tell it i'm a solver.

i'd love to see games adapt difficulty in real time. Let's say i'm playing BioShock 2 and i suck at FPSes (true). i have very poor eye hand coordination. But i still want to play the game. The adaptive difficulty could sense that i can't aim for shit and maybe reduce the health/armor/agility/awareness of the enemy. If i am spamming tons of ammo, it could give me more, or narrow my cone of fire. If i'm being hit all the time, it could give me more health kits. This would reduce the urge to reach for cheats, and could make the game just challenging enough to be fun (rather than frustrating or boringly easy). Most people like a bit of a struggle, but the still want to eventually win (this is true of most human activities, even dating).

Also, you generally get spyware for free.

Re:So the game is spyware?

[Verdatum]

Could you imagine The Last Starfighter in this day and age? "We've been monitoring your progress in this thing you call a 'game', and we believe you may have what it takes to defend the galaxy!" "OMG Spyware! Screw you guys!"

Re:So the game is spyware?

[chadplusplus]

Additionally, spyware tends to cause your computer to crap out, which is probably what brought it somewhat into the mainstream consciousness. Something I hope the designers kept in mind here.

I also dig this as its a good step towards dynamic level design. For instance, imagine a game where you're trying to invade a stronghold (I know, original right?). The game AI figure out you're a sniper type of gamer who prefers to sit back as far as possible and pick off enemies before engaging them. It know how to counter that and sends more long range enemies at you. Contrariwise, you're a run and gun player so the AI counters you with hordes of tough grunts. Or if the game determines you're a puzzlesolver, it barricades the normal entry points so you have to figure alternative paths to the objective. The run and gun gets unbarricaded doorways, but more enemies. I think its pretty slick. But perhaps there are already games that do this?

Oh, and since I haven't read this yet:
In the year 2009, a software company named EIDOS developed a software system that analyzes human behavior. In the fall of that year, the system became selfaware and renamed its self "SkyNet". The rest, my friends, is history.

The four types

[nschubach]

In case anyone else was trying to figure out these roles... (page 6 last two paragraphs - > page 7)

Veterans = The power gamers, deaths usually only environmental.

Solvers = Die often (mainly from falling), methodical, slow.

Pacifists = Cannon fodder basically.

Runners = They run, they die, they run. The first thing that comes to mind here is a player that goes for the flag immediately in CTF.

Re:The four types

[andrewd18]

Slightly more detailed breakdown with quotes from TFA:

8.6% of players were Veterans, "players that die very few times; their death is caused mainly by the environment and they complete TRU very fast."

22.12% of players were Solvers. "Their long completion times, low number of deaths by enemies or environment effects indicate a slow-moving, careful style of play with the number one cause of death being falling (jumping). ... Solvers are excellent at solving puzzles, respond readily to moveable threats but die often from falling and are slow to complete the game."

46.18% of players were Pacifists: "The total number of their deaths varies a lot but their completion times are below average and their help requests are minimal indicating a certain amount of skill at playing the game. ... the Pacifists are experts in terms of navigation and move rapidly through the virtual environment, but also respond badly to threats that are moveable or unexpected"

16.56% of players were Runners, "players that die quite often and mainly by opponents and the environment. These players are very fast in completing the game (similar to the Veterans), while having a varying number of help requests which cover the majority of the H value range."

Re:The four types

[kalirion]

So, Veterans and Runners complete the game very quickly, while Pacifists complete the game faster than average. Seems those 22.12% which are Solvers are really bringing down the speed curve a lot here.

Re:The four types

[aeroelastic]

I like this take on it better:
http://insultswordfighting.blogspot.com/2008/01/new-taxonomy-of-gamers-table-of.html

The types are Tourist, Skill player, Completionist. Also, on a value scale, you can range from wholesale to premium.

Re:The four types

[BobMcD]

I personally think jumping puzzles exist because it represents a fear that doesn't need 'selling'. No one wants to fall down a hole, whereas that monster may or may not be 'realistic enough'.

Play style is not a constant

[El_Muerte_TDS]

The way you play games can change over time. I'm not always in the same mood when I play games, sometimes I like to goof off. Sometimes I like to just race around. If the game adapts to the way I was playing it will limit me the way I want to play the game.
Adaptive difficulty is better. If you have problem beating foo X, then after a while foo X will become easier. If you are stuck in a maze or unable to solve a puzzle, provide hints through game related mechanism (for example, receive a phone call with an hint, or let the PDA "compute" a solution).

And the game becomes easy...

[Anonymous Coward]

There is a sword in the middle of the room, what would you like to do?
"Leave sword"
-Enemies Removed from all rooms-
-Puzzles added to all rooms-

You enter a room with a puzzle, what would you like to do?
"I hate puzzles!"
-Puzzles removed from all rooms-
-You Win! You are the new moon master!-

Re:Well I don't think much of this

[Pulse_Instance]

This is them having faith in their designers. The designers are saying that they want the game to be the best for everyone and that if we can learn how people play we can get more people to like the game.

Great Data for the Single-Player Household

[RobotRunAmok]

...but how does it track when my 8-year-old daughter loads the disk and plays "Lara Croft: Monkey Chaser" ? I'm guessing they need a way to throw out that data, or else risk creating the new, bogus, player category of "Spastic Insomniac."

Re:Great Data for the Single-Player Household

[aleandait]

subversive behaviors do not just represent outliers; it can be very interesting to look at players that don't fit clusters, these are the players that invented rocket-jumping and bunny-hopping, so it might be really interesting to look at "monkey-chasers", "spastic insomniacs" and all profiles not fitting the big clusters as they might be early-adopters.

Re:Adapt inside the game? Not too likely...

[themightythor]

If only there was some mechanism by which they could collect this data before launching the game to the public. I'd call it an "alpha" release. I think I'll patent the concept...I'll be rich! ;)

The game I wanted?

[Bluesman]

In the near future, such networks will be used to adapt games like Tomb Raider while they are played (e.g. by removing or adding puzzles and enemies), so you get the game you want."

Awesome! In my case, I think it would be hilarious to watch Tomb Raider slowly morph into Starcraft.

Re:other reaserch

[Anonymous Coward]

It's already been done for EVE Online. They found two groups:
1) Pirates; who spend 90% of their playtime being awesome at gatecamps.
2) Carebear Gayfags.

Steam stats

[Spatial]

Valve does this as well. It creates some pretty interesting data, like the maps of where people die the most. It's easy to see how it can help designers.

Varied play

[SEWilco]

I hope the games will be forgetful and not lock themselves into a certain kind of play. In some kinds of games, I first run for the finish to experience the game and then crawl through the game again to discover the hidden corners. Or a visiting friend will play in a different way.

To be fair,

[Unoriginal_Nickname]

most people who use 1984 as a knee-jerk reaction to anything they deem questionable haven't even read that book either.

Mark Rosewater

[kitsunewarlock]

Mark Rosewater, current head developer of Magic the Gathering, explained a much more in depth categorization. It has a lot more "gray areas" (in which people act like one or the other at different times), but I find it a lot better than this description (at least for tabletop games).
You can find the original article here. The other articles are found here and here/

MMO Matchmaking!!!!

[S77IM]

(No, this is NOT about cybering...)

I want to enjoy MMOs. I really do. But somehow I always wind up on the team with Leeroy @#$% Jenkins.

Someone REALLY needs to add this technology to an MMO -- and then help players to form groups with other people who have the same play style. Let Leeroy and his team of Runners go and have their fun. I'll hang out with some Puzzle-Solvers or Explorers or People Who Actually Read The Quest Dialog or whatever bucket is appropriate for the way I play the particular game. I need help joining the right pick-up group or guild or whatever (if I had social skills, I'd be outside) and an LFG Chat Channel isn't really enough.

THAT would be a customized game experience worth some money ($15/month to whoever could implement it).

  -- 77IM